The DOI matches its Crossref record; 50 of the 50 references it cites were found in a registry or an index.
Semantics derived automatically from language corpora contain human-like biases
Open accessAbstract
Machines learn what people know implicitly: AlphaGo has demonstrated that a machine can learn how to do things that people spend many years of concentrated study learning, and it can rapidly learn how to do them better than any human can. Caliskan et al. now show that machines can learn word associations from written texts and that these associations mirror those learned by humans, as measured by the Implicit Association Test (IAT) (see the Perspective by Greenwald). Why does this matter? Because the IAT has predictive value in uncovering the association between concepts, such as pleasantness and flowers or unpleasantness and insects. It can also tease out attitudes and beliefs—for example, associations between female names and family or male names and career. Such biases may not be expressed explicitly, yet they can prove influential in behavior. Science , this issue p. 183 ; see also p. 133
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Training deep learning models on biased data can amplify human biases, replicate toxic language, or perpetuate stereotypes of gender, sexual orientation, political leaning, or religion (e.g., Caliskan et al. 2017; Hartmann et al. 2023b).”
From Generative AI (2024) · cites this paper as Caliskan et al. 2017
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Some studies (Caliskan et al. 2017;Greenwald et al. 1998) have found that names associated with European Americans are more likely to be linked with pleasantness, while non-European names tend to be associated with unpleasantness.”
From Towards trustworthy LLMs: a review on debiasing and dehallucinating in large language models (2024) · cites this paper as Caliskan et al. 2017
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Machine learning models reproduce various content biases inherent to their training data, including gender bias, racial or ethnic bias, bias for negative and threat-related information, and socio-economic bias 69,[86][87][88][89][90] .”
From Machine culture (2023) · cites this paper as [87]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“If people type “He is a doctor. She is a nurse” into Google Translate and translate the phrases into Turkish, Google Translate provides: “O bir hemşire. O bir doktor”. Those Turkish sentences are gender-neutral; Turkish does not differentiate between the words “he” and “she”. When translating the Turkish text into English again, Google Translate provides: “She is a nurse. He is a doctor”.”
From Discrimination, artificial intelligence, and algorithmic decision-making (2025) · cites this paper as Caliskan, Bryson and Narayanan 2017
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“That most models produced overgeneralizations reliably across time suggests that they may have incorporated an ‘algorithmic bias’ from the training data similar to the way models trained on linguistic data can acquire gender, racial, or political biases, by picking up human tendencies implicit in training text corpora [51–53].”
From Generalization bias in large language model summarization of scientific research (2025) · cites this paper as [51]
Experimental — accuracy not yet measured
13 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Glove: Global Vectors for Word Representation
Pennington, J., Socher, R., & Manning, C. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1162 · Cited by 34,071
Efficient Estimation of Word Representations in Vector Space
Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1301.3781 · Cited by 26,568
Distributed Representations of Words and Phrases and their Compositionality
Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1310.4546 · Cited by 24,533 · cited 3×
“For example, we repeated all the WEAT and WEFAT experiments presented above using a different pre-trained embedding: word2vec on a Google News corpus (Mikolov and Dean, 2013).”
BackgroundMethodMeasuring individual differences in implicit cognition: the implicit association test
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.74.6.1464 · Cited by 9,210
Bertrand, M., & Mullainathan, S. (2004) · American Economic Review · doi:10.1257/0002828042002561 · Cited by 4,504
Dwork, C., Hardt, M., Pitassi, T., Reingold, O., & Zemel, R. (2012) · Proceedings of the 3rd Innovations in Theoretical Computer Science Conference · doi:10.1145/2090236.2090255 · Cited by 3,501
A Neural Probabilistic Language Model
Bengio, Y., Ducharme, R., Vincent, P., Jauvin, C., Jaz, K., Hofmann, T., Poggio, T., & Shawe-Taylor, J. S. (2000) · Cited by 3,479
Barocas, S., & Selbst, A. D. (2014) · doi:10.2139/ssrn.2477899 · Cited by 3,048
From Frequency to Meaning: Vector Space Models of Semantics
Turney, P. D., & Pantel, P. (2010) · Journal of Artificial Intelligence Research · doi:10.1613/jair.2934 · Cited by 2,866
Pattern recognition and machine learning
(2007) · Choice Reviews Online · doi:10.5860/choice.44-5091 · Cited by 2,650
Research works printed but not linked: 1
- M. Stubbs Text and Corpus Analysis: Computer-Assisted Studies of Language and Culture (Blackwell Oxford 1996).
Cited by
Dissecting racial bias in an algorithm used to manage the health of populations
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019) · Science · doi:10.1126/science.aax2342 · Cited by 7,137
A Survey on Bias and Fairness in Machine Learning
Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2022) · ACM Computing Surveys · doi:10.1145/3457607 · Cited by 5,812
Training language models to follow instructions with human feedback
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2203.02155 · Cited by 4,208
The Kinetics Human Action Video Dataset
Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., Natsev, P., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.06950 · Cited by 2,593
Explaining Explanations: An Overview of Interpretability of Machine Learning
Gilpin, L. H., Bau, D., Yuan, B. Z., Bajwa, A., Specter, M., & Kagal, L. (2018) · 2018 IEEE 5th International Conference on Data Science and Advanced Analytics (DSAA) · doi:10.1109/dsaa.2018.00018 · Cited by 2,507
“…to Explanation” [12], calls for diversity and inclusion in AI systems [13], findings that some automated systems may reinforce inequality and bias [14], and requirements for safe and secure AI in safety-critical tasks [15], there has been a recent explosion of interest in interpreting the…”
Background
Metadata
- Authors
- Caliskan, A., Bryson, J. J., & Narayanan, A.
- Affiliations
- Princeton University (US)
- Corresponding author
- Caliskan, Aylin
- Bryson, Joanna J.
- Narayanan, Arvind
- Venue
- Science
- Published
- 2017-04-14
- Volume
- 356
- Issue
- 6334
- Pages
- 183-186
- ISSN
- 0036-8075 (ISSN-L)
- 1095-9203
- OpenAlex
- W2893425640
- Cited by
- 2,887
- References
- 50
- Citations by year
- 2026: 299 · 2025: 368 · 2024: 343 · 2023: 411 · 2022: 326 · 2021: 399 · 2020: 326 · 2019: 224 · 2018: 134 · 2017: 47
- FWCI
- 459.41
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Science (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- The University of Bath Online Publications Store (The University of Bath) (submittedVersion)
- Pure (University of Bath) (submittedVersion, other-oa)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Sociology and Political Science › Social and Intergroup Psychology
- Physical Sciences › Computer Science › Artificial Intelligence › Sentiment Analysis and Opinion Mining
- Keywords
- semantic bias, machine learning, text corpora, Implicit Association Test, algorithmic bias, natural language processing, gender bias, racial bias, cultural bias, bias detection
- MeSH terms
- Machine Learning, Association, Female, Humans, Language, Male, Names, Semantics, Sex Factors, Internet
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Semantics derived automatically from language corpora contain human-like biases. https://app.cochat.ai/papers/10.1126/science.aal4230
- The paper
- Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora contain human-like biases. Science. https://doi.org/10.1126/science.aal4230
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.